PROJECT SUMMARY/ABSTRACT Cardiac arrest secondary to hemorrhage accounts for up to 51% percent of trauma deaths and is the leading cause of preventable death after penetrating injury. There are no known risk factors for mortality in cardiac arrest after injury that are subject to modification by providers or hospitals. Research into the quality of resuscitation in non-traumatic cardiac arrest has demonstrated that higher quality cardiopulmonary resuscitation is associated with improved survival, thus elucidating targets for interventions designed to improve outcomes. No analogous investigations into the quality of resuscitation exist in patients with cardiac arrest secondary to injury, where Resuscitative Thoracotomy (RT) is the resuscitative procedure of choice. One barrier to such inquiry is data collection. Resuscitative thoracotomy is an emergent high-impact low- frequency event. Prospective data collection requires continuously present and available research personnel, a resource that is beyond the reach of most centers, whereas retrospective data collection using chart review is challenged by incomplete capture of important data elements. The long-term goal of our research is to reduce mortality secondary to cardiac arrest after injury by identifying of modifiable risk factors for mortality, trialing interventions based on these risk factors, and developing best practices based on these trials. In this pilot proposal, we will leverage an existing audiovisual recording system at a level I trauma center with a high volume of patients presenting in cardiac arrest after injury to study the quality of RT. Our central hypothesis is that increased RT speed is associated with higher rates of return of spontaneous circulation (ROSC) after cardiac arrest secondary to injury. We also hypothesize that for emergent high-impact low-frequency events such as RT, data collection using audiovisual recordings will result in less bias, increased precision, and decreased missing data compared to more established methods such as prospective real-time of data collection and data collection using retrospective chart review. In specific aim 1, we will use data collected from a continuously running audiovisual recording system to determine the association between speed of RT (as defined by times to completion of procedural milestones) and rates of ROSC. In specific aim 2, we will demonstrate that compared to prospective real-time data collection or retrospective chart review by trained research assistants, data collection using abstraction of audiovisual recordings of HILF events results in decreased bias, increased precision, and decreased rates of missing data.